RXTE Helpdesk/FAQ RXTE What's New HEASARC Site Map


RXTE
GOF
Using fasebin to Create Phase-Binned Spectra and Histograms
Recipes from the RXTE Cook Book
RXTE
FAQ


Introduction

The recommended tool for producing phase-binned spectra is fasebin. Fasebin handles the extraction, barycenter correction, and phase-binning of input RXTE data, eliminating the need to run separate tools for these various functions. It produces a single FITS file in the OGIP Type 2 spectral format: each row in the output file is a separate spectrum associated with a given phase bin. Other tools in the fasebin suite can then be used to subtract persistent emission (fbsub), combine phase-bins (fbfsum), and produce a "counts-per-phase-bin" histogram (fbssum). Examples of each will be shown below.


Data files you will need

To run fasebin you will need to have the following:


Running fasebin

  1. For users analyzing data in GOOD XENON MODE ONLY:

    Run make_se on your data files, to produce full binary FITS files from the Good_Xenon1 and Good_Xenon2 raw data files. Make an ascii list of these new files, and treat this as your list of PCA data files in all future steps.

  2. Determine which PCA or HEXTE data files correspond to which Orbit Ephemeris files (ie., which observations occurred on which days). Segregate these files into separate ascii lists by orbit day. The end result should be a separate file list for each orbit day, with each list giving only the data files which occurred on that day. (XDF can be a useful tool for matching observations to Orbit files; see the RXTE Getting Started Guide for details on using XDF.)

  3. Look at the text file:

    $LHEA_DATA/psrtime.dat

    (This directory and file are created automatically when FTOOLS is installed on your system.)

    This file contains a recent copy of the Princeton Radio Pulsar database. If your source is in this file with the correct epoch and period, you may skip steps #4 and 5, and go directly to step 6. (Obviously, if your source is not a radio pulsar it will not be in this file; proceed directly to step #4.)

  4. Create a new directory (eg., "fasebin_files") and type the following UNIX command:

    setenv TIMING_DIR /full/path/name/fasebin_files

    where "/full/path/name/fasebin_files" is the full UNIX path name of the directory you just created.

  5. Copy the psrtime.dat file into your new directory and edit it to include an entry for the source you wish to analyze. Eg.:

    cp $LHEA_DATA/psrtime.dat $TIMING_DIR/psrtime.dat

    edit $TIMING_DIR/psrtime.dat

    • source entries are in alphabetical order by RA-Dec name (first column, labled "PSR B")

    • create a new entry for your source by making a source name from its RA and Dec in hexadecimal coordinates; place the entry in the proper location to preserve the ascending order of source names.

    • fill in the RA and Dec positions of your source (in J2000)

    • place the period in inverse seconds in the "f0(s^-1)" column

    • the "MJD1" and "MJD2" columns contain the time span in MJD over which this period is valid. Fill in appropriate values, but make sure this time span includes the dates of the RXTE observations you wish to analyze.

    • the Web tool xTime can be used to convert times into MJD (xTime is also found under the "Proposal and Planning Tools" button on the GOF homepage).

    • place the reference time for phase=0 in the "MJD(t0)" column. Note this time must be in MJD, and it need not fall within the time span specified in "MJD1","MJD2".

    • (optional) if you have a first or second derivative for the period, or an RMS error, fill these in under the appropriate columns

    • place the source name again in the final column ("PSR J")

    • save and exit file

  6. If your source has a binary companion and you would like to correct for its orbit, open the following file with any text editor:

    $LHEA_DATA/psrbin.dat

    This is the binary orbits table for the Princeton Radio Pulsar Database. If your source is listed (under "PSR B" column), proceed to step 7. If not, create a $TIMING_DIR as in step #4 (unless you have already done so), and follow steps #4 and #5 for file "psrbin.dat". Place the appropriate binary orbit information in your copy of the psrbin.dat file instead of the pulsar period information listed in step #5. (See "fhelp fasebin" for more information on the data columns in the Princeton Radio Pulsar Database binary tables.)

  7. You're finally ready to run fasebin:

    given:

      orbit file: FPorbit_Day1107
      list of PCA data files: binned_data_day1107.txt
      GTI file spanning these data: clean_data.gti

    type:

        fasebin
    
        Orbit File[] FPorbit_day1107
        Name of output .pha file[faseBin.pha] day1107.pha
        Name of GTI file[none] clean_data.gti
        Source Name[] 0833-45
        Number of phase bins[100] 20
        Datafile(s)[] @binned_data_day1107.txt
    

    This will produce an output file called "day1107.pha" with spectra in 20 phase bins for source "0833-45", whose period and other characteristics were read from the "psrtime.dat" and "psrbin.dat" files.

    N.B.: fasebin produces spectra for 2 full phase cycles. Thus, if you requested N phase bins, your output file will contain 2N+1 spectra covering two full phase cycles, with the second cycle being identical to the first.

    Good Xenon data only: fasebin automatically extracts data from only the first (top) layer of the PCA when Good Xenon data are used. For faint sources (~ < 20 cnt/sec/PCU) which don't have unusually hard spectra, this greatly increases the signal-to-noise of the accumulated spectra. If you would rather collect data from the full PCA (all 3 layers) use this hidden parameter on the command line when you invoke fasebin:

              fasebin l1only=no
    
    See "fhelp fasebin" for further details.


Combining fasebin files from different Orbit days

To combine data of the same source and configuration taken over multiple Orbit days:

  • repeat step #7 for each Orbit day, until all data from all days has been phase-binned

  • create an ascii file list with the names of the all the output phase-binned spectral files, eg:

     ls day*.pha > alldays_pha.list
     more alldays_pha.list
     day1107.pha
     day1108.pha
     day1112.pha
    
  • run fbadd to make a single phase-binned spectral file combining data from all these days:

     fbadd
     Input File (or @filelist)[] @alldays_pha.list
     Name of output .pha file[fBadd.pha] alldays.pha
    


Analyzing fasebin spectra in xspec

Phase-binned spectral files from any of the fasebin tools may be read directly into xspec 10.0 or higher. (Earlier versions of XSPEC do not support OGIP Type 2 Spectral FITS files.) The syntax for reading Type 2 spectral files in xspec is:

 XSPEC>  data alldays.pha{10}

This will read the 10th spectrum (ie., the tenth row) of your phase-binned .pha file; to display more than one spectrum at a time, use the "data 1", "data 2" syntax in xspec (see the xspec manual for details).

To produce a response matrix, DO NOT give your phase-binned spectral file to pcarsp. Instead, extract a normal, un-phase-binned spectrum for the same type of data you gave fasebin (eg., if you ran fasebin on PCA binned mode data, run saextrct on some part of the same binned data and produce a spectrum). Give this un-phase-binned spectrum to pcarsp and make a response matrix from this. Since you are using the same data that you gave to fasebin, the response matrix will be valid for the fasebin spectral file as well. (Good Xenon users only: keep in mind that fasebin, by default, extracts only layer 1. If you used this default behavior, make sure you specify layer 1 only (LR1) when you run pcarsp.)


Subtracting persistent emission

You may subtract the persistent emission (the baseline background) from your data in one easy step, with the FTOOL fbsub:

 fbsub
 Input File[] alldays.pha
 Name of output .pha file[fBsub.pha] alldays_pulsed-only.pha
 Phase minimum[0.0]
 Phase maximum[1.0]

The phase minimum and phase maximum may contain any continuous range of phase space which you feel best represents the unpulsed component for your source. The entire phase range (0.0 to 1.0) may be used in most cases as a good first approximation, since the few phase bins containing the pulsed emission will not effect the baseline much. You may also enter a second phase range to be used in calculating the unpulsed background -- see "fhelp fbsub" for details.


Combining several phase bins into one

You may combine contiguous phase bins into a single spectrum using the FTOOL fbsum:

    fbfsum binstart=5 binstop=8
    Input File[] alldays.pha
    Name of output .pha file[fBfsum.pha] alldays_phase5-8.pha
    Number of phase bins (rows) to be averaged[5] 4

This will create a Type 2 spectral file with a single row representing the combined spectrum for phase bins 5-8. If the last input (number of phase bins) was changed from "4" to "2", the output file would contain two spectra, one for phase bins 5+6, the other for bins 7+8. Leaving out the "binstart=5 binstop=8" parameters on the command line would combine every 4 phase bins in the input file into a single spectrum in the output, resulting in an output file with 5 separate spectra. (The input file contains 20 phase bins.)


Creating a counts-per-phase-bin histogram

To produce a histogram of counts-per-phase-bin, use the FTOOL fbssum:

    fbssum
    Input File[] alldays.pha
    Channels for energy ranges[]

This will sum the counts in all channels for each phase bin, and dump the resulting totals to the screen. Header-style comments at the top of the output define the columns and units, as well as echoing the included channels. You may cut and paste the entire output to a file and plot with any ascii plotting package (eg., QDP, gnuplot, etc.).

Please Note: the resultant histogram covers two full phase cycles, ie., 0.0 to 2.0 in phase space. Thus, if you requested N phased bins when running fasebin, you will have 2N+1 phase bins in your output.

For the syntax on creating histograms for restricted channel ranges, see "fhelp fbssum". BEWARE: the channels are relative to the configuration used, NOT THE ABSOLUTE 0 to 256 CHANNEL RANGE OF THE PCA INSTRUMENT. See the fhelp for further details.


If you have a question about RXTE, please send email to one of our
help desks.

This page is maintained by the RXTE GOF and was last modified on Thursday, 16-Sep-1999 08:38:13 EDT.